Abstract
Background
Most cohort-based research for subfertility has been conducted in clinic-based cohorts, which may differ from population-based cohorts.
Methods
We retrospectively recruited parallel cohorts of subfertile women: one by sampling two specialty fertility clinics in Utah, and one by population-based sampling based on marriage and birth records. The index date (of first clinic visit or subfertility status) was between 2000 and 2009, and we linked the women recruited to subsequent birth certificate records through December 2010.
Results
We enrolled 459 women through clinic-based sampling and 501 women through population-based sampling. Clinic-based women were older, had higher annual household income and more likely to have had a most intensive treatment of intrauterine insemination (31%) or in vitro fertilisation (46%) than women from population recruitment (19% and 14%, respectively). Conversely, they were less likely to have received no medical treatment (9%) compared to women from population recruitment (41%). For both types of sampling, prior to eligibility screening, non-responders were less likely to link to a live birth than responders: 51% versus 58% for clinic-based, and 69% versus 76% for the population-based with an index date in 2004.
Conclusions
Population-based sampling for subfertility cohort research identifies women who were more likely to have had less intensive treatment or no treatment. However, in both clinic-based and population-based sampling, women who have had a live birth are more likely to respond to retrospective recruitment.
Keywords: retrospective cohort, subfertility, population-based, clinic-based
Introduction
Subfertility, also called infertility, is traditionally defined as failure to achieve pregnancy after 12 or more months of regular unprotected intercourse.1, 2 An estimated 10–15% of women or couples experience subfertility at some point in their reproductive life and about 15% of U.S. women who are currently attempting to become pregnant are experiencing difficulty becoming pregnant.3, 4 About half of these women or couples seek medical treatment, with significant disparities among socioeconomic and ethnic groups.3, 5, 6 Subfertile couples and their clinicians must choose among a variety of treatment options, including no medical treatment and/or alternative therapies.2, 7 Outcomes data needed to inform these choices include the cumulative live birth rates and the health outcomes of pregnancies, newborns and children.8–11
Most of the data available for reproductive outcomes of fertility treatments are collected from patients enrolled from specialty fertility clinics, but many couples seek treatment from other types of providers, or do not seek treatment at all.5, 7, 9, 12 Similarly, many studies focus on outcomes for treatment with in-vitro fertilization (IVF), with or without intracytoplasmic sperm injection, but fewer studies address outcomes for non-IVF fertility treatments.8 Inferences about effects associated with specific treatments require appropriate comparisons to other subfertile couples receiving other treatments, or no treatment.9, 11, 13, 14
To address these gaps, we conducted an observational retrospective cohort study with parallel clinic-based and population-based cohorts of women residing in the State of Utah, who reported a history of primary subfertility (i.e., trying to conceive for at least one year, with no prior pregnancies). The clinic-based cohort was recruited from the two major specialty fertility clinics in Utah during the time of the study; while the population-based cohort of women was recruited based on sampling from state marriage and birth records. The study has two primary aims: 1) to compare the cumulative proportion of women experiencing live birth and the time to live birth by type of fertility treatment chosen (IVF, IUI, OD, or no medical treatment); 2) to examine perinatal outcomes, including pre-term birth, low birth weight, birth defects, and neonatal morbidity for children conceived through these different treatments.
This paper reports on the design of the Fertility Experiences Study, and the recruitment, enrollment, and characteristics of those enrolling clinic-based and population-based cohorts. We hypothesized that there would be substantial differences in baseline reproductive history between subfertile women recruited from fertility clinics and geographically and age-matched subfertile women recruited directly from the reference population. We further hypothesized that there would be substantial differences in treatments received after the index date. We also explored whether there was any evidence of differential response to the invitation to be screened between the clinic and population sampling.
Methods
Design, target populations, and time frame
The Fertility Experiences Study consists of two parallel retrospective cohorts. For the clinic-based cohort, we sought to enroll women seen for an initial visit for primary subfertility at one of the two participating fertility specialty clinics during specific time frames; the Utah Center for Reproductive Medicine (UCRM), affiliated with the University of Utah; or the Reproductive Care Center (RCC), a private practice. Both practices are located in the Salt Lake City metropolitan area and serve a referral base that includes the entire State of Utah. For the population-based cohort, we sought women from the entire State of Utah who had also experienced primary subfertility in parallel time frames. We planned the initial index date for first clinic visit in 2004 for UCRM, with a comparable date of assessing primary subfertility for the population cohort (December 31, 2004). After assessing our initial rates of contact and seeking to reach our planned sample size (n=1000), we added a second wave of recruitment with an index date in 2008 for UCRM and the population cohort. About the same time, we also added recruitment for RCC, with initial clinic dates spanning 2000–2009 (the full span of years was more logistically practical for RCC).
Preliminary eligibility
All potential participants had to meet initial eligibility criteria prior to recruitment, shown in Table 1. Initial eligibility criteria were designed to identify women most likely to meet the final eligibility criteria, based on information readily available at each clinic, or for the population sample, the Utah Population Database (UPDB).15 The UPDB linked marriage and divorce records to birth certificates, fetal death certificates and adoption certificates. The UPDB generated a random sample of women meeting preliminary eligibility criteria, to be contacted by the authorized independent intermediary agency for recruiting participants from the UPDB, called the Resource for Genetic and Epidemiologic research (RGE).
Table 1.
Eligibility criteria
| Clinic cohort preliminary criteriaa |
| Female |
| New patient for infertility at UCRM in 2004 or 2008; or RCC between 2000–2009 |
| Male partner |
| Age 20–35 at first visit (UCRM only) |
| No known pregnancies prior to first appointment at clinic (UCRM only) |
|
Population cohort preliminary criteriab consists of two period subcohorts (a) and (b) |
| Female |
| Married between (a) January 1, 2000 and December 31, 2002, or (b) between January 1, 2004 and December 31, 2006 |
| Age 18–30 at date of marriage |
| Married 2–5 years to same person as of index datec |
| Husband living as of index datec |
| No live births or fetal deaths recorded in Utah as of index datec |
| Current Utah address within the past 5 years |
| Final criteria for both cohortsd |
| Female Age 20–35 at index datec Clinic cohort: date of first clinic visit Population cohort: (a) December 31, 2004, or (b) December 31, 2008 |
| No pregnancies prior to index datec |
| At least one year of trying to get pregnant with male partner at index datec |
| Residence in Utah during entire three years following index datec |
based on information from clinic records
based on information from the Utah Population Database, a random sample of those eligible was contacted
Clinic cohort: date of first clinic visit (UCRM in 2004 or 2008; RCC between 2000–2009); population subcohort (a) December 31, 2004; population subcohort (b) December 31, 2008
based on information reported by the women during online or telephone screening
Recruitment and screening
All potential participants received a letter by mail, explaining the study and inviting them to respond regarding their interest. Potential participants in the clinic cohort received a letter from their clinic, signed by the director of the clinic and the principal investigator for the study (JBS). This letter informed them that the study was for “women who have experienced difficulties with fertility,” and that the study would address “treatment choices, outcomes, and costs that couples experience.” Response options included immediately completing the screening questionnaire (by phone or web form), requesting further information, or requesting not to be contacted further about the study. Study staff attempted to make follow up telephone calls for women from UCRM. No telephone calls were made for women from RCC.
Potential participants in the population cohort received a mailed letter from RGE explaining the study. This letter informed them the study was about women’s fertility and “would involve completion of a survey about any times in your life you may have tried to get pregnant, and care you have received, if any.” Response options included requesting their contact information to be given to study personnel, or requesting not to be contacted further about the study. For women interested, their contact information was given to study staff, who subsequently contacted participants to invite them to complete the screening questionnaire by phone or web form. RGE staff also attempted to contact by telephone those women who did not respond to the initial letter from RGE.
Final eligibility, enrollment, and consent
Final eligibility criteria were the same for both the clinic and population cohorts, and were based on information reported by the women in the screening questionnaire, as shown in Table 1. All potential participants completed the screening questionnaire by telephone or a web form, using Opinio software (ObjectPlanet, Inc., Oslo, Norway). The screening questionnaire took less than 5 minutes to complete.
Women who met final eligibility criteria based on the screening questionnaire were invited to participate in the study, usually immediately. In the first screens of the online questionnaire, the study procedures were explained, and women were informed that completing the questionnaire constituted their informed consent and enrollment in the study. The study was reviewed and approved by the University of Utah Institutional Review Board for research involving human subjects.
Duplicate participants between cohorts
Duplicates between cohorts were identified during screening, or in a few cases, subsequently during data analysis by name, birthdate, and/or email address (n=5). These 5 women were included only in the clinic cohort. Although some women reported they had received care at both UCRM and RCC, they responded to only one clinic invitation.
Data Collection
Fertility Experiences Questionnaire
Participation in this study consisted of completing a mixed-mode, two-part questionnaire, the Fertility Experiences Questionnaire (FEQ). The first phase of the FEQ was administered via the internet, using Opinio software (ObjectPlanet, Inc., http://www.objectplanet.com/opinio/), and required 25–45 minutes to complete. (Four women completed a paper questionnaire by mail.) The second phase was an in-depth phone interview conducted by trained study staff, requiring between 20 and 45 minutes to complete. During this interview, more detailed questions reassessed screening criteria, but women’s eligibility was not altered by any responses at this stage. We have reported elsewhere an initial validation of several components of the FEQ, the sources used to develop it, and its contents.16
Data linkage
After study data collection was complete, we mailed a letter to study participants informing them that we planned to link their study data to Utah vital records including birth certificates and fetal death certificates, with an option for participants to notify us if they wished to opt out of linkage. Seven women opted out of linkage. In addition to this linkage, RGE created a de-identified files of data from birth certificate and fetal death certificates for all UCRM clinic patients, as well as from the population sample who did not respond to the invitation letter about the study.
Compensation
Women who completed both the online and telephone portions of the Fertility Experiences Questionnaire were sent a $10 online gift certificate.
Outcome measures
The main outcome measures for the study are time to pregnancy, pregnancy, live birth, and additional pregnancy outcomes including spontaneous abortion, premature birth, low birth weight, and birth defects. These outcomes are ascertained by linkage to birth certificates, fetal death certificates, data from the Utah Birth Defects Network, and by responses to the FEQ. Outcomes for this report include proportion of women responding, eligible, and enrolling, with associated demographic and reproductive characteristics at baseline, and also subsequent live birth, as assessed by the FEQ, clinic records, and linked birth certificates.
Sample size
Based on assumptions of distributions of treatment choices and associated cumulative pregnancy rates in the clinic and population cohorts, we planned to enroll 1000 women, with 500 from the clinic-based cohort and 500 from the population-based cohort in order to have sufficient power to assess differences between IVF and other treatments in cumulative proportion achieving live birth, preterm birth, and low birth weight (among all pregnancies).
We attempted to contact all potentially eligible women first seen at UCRM in 2004, and all of the potentially eligible women first seen at RCC in 2000–2009. As we approached the targeted sample size for the clinic cohort, we attempted contact with a random subset of women first seen at UCRM in 2008. For population recruitment, a random subset of those meeting preliminary eligibility criteria in the UPDB was contacted to keep population recruitment approximately equal to clinic recruitment.
Statistical analysis
Descriptive statistics were used to summarize demographic and reproductive characteristics of nonrespondents (where available), respondents, and participants. The chi-square and t-test statistics were calculated to assess differences between nonrespondents and respondents, between clinic and population cohorts overall, and for the subgroups of women who never received medical treatment.
Results
A total of 26,678 letters was sent throughout the study: 10,677 for clinic recruitment and 16,001 for population recruitment. Subsequently, 459 women were enrolled in the clinic group and 501 women in the population group. The vast majority also completed the follow-up telephone interview component of the FEQ. Figures 1–2 show the recruitment flow for the clinic and population cohorts. While 43% of the clinical women screened eligible, compared to 31% of the population women screened, essentially the same proportion of those eligible in both groups enrolled (87% and 88%, respectively). Supplementary Figures 1–5 give full recruitment details for each of the subgroups, i.e., the UCRM 2004, UCRM 2008 and RCC 2000–2009 clinic groups, as well as the UPDB 2004 and UPDB 2008 groups. Overall rates of interest varied from 11% (RCC) to 27% (population group with an index date of December 31, 2008). Compared to the other subgroups, the proportion of interested women who were screened in the UCRM 2008 subgroup was lower (78%), because of the shorter time frame that recruitment and screening were active for this subgroup (3 months).
Figure 1.
Recruitment, screening and enrollment in the clinical cohort
awomen could have more than one reason for initial ineligibility or non-confirmed eligibility
Figure 2.
Recruitment, screening and enrollment in the population cohort
a6230 letters were returned with an incorrect postal address.
bwomen could have more than one reason for initial ineligibility or non-confirmed eligibility
Comparative Description of the Clinic and Population Cohorts
Women recruited from the clinic were older on average than those from the population (p<0.001). Most women in both cohorts (93% and 95%, respectively) identified as white, non-Hispanic; and just over 60% of each cohort had graduated from college. Women in the clinic cohort had somewhat longer attempts to conceive, with or without subsequent pregnancy and regardless of use or timing of fertility treatment (43% over 60 months) than women in the population cohort (38% over 60 months; p=0.068), and substantially longer time between the start of their first attempt to conceive and the study telephone interview (29% had 10 years or more versus 7%, respectively; p<0.001). Using mutually exclusive categories of most invasive treatments received: in the clinic cohort, only 9% had never received any type of medical fertility treatment, 14% had used fertility drugs, 31% had used artificial insemination, and 46% had used IVF; the respective proportions for the population cohort were 41%, 26%, 19%, and 14%, respectively (p<0.001). Details for both cohorts are given in Table 2, and for each separate clinic and population subgroup in Supplemental Table 1. We also compared women who never received medical fertility treatment who were recruited either from the clinic (n=43), or from the population (n=207). Those from the clinic were more likely to be older: 24% age 31 or more, versus 4% from the population group (p<0.001). However, there were no significant differences between these two subgroups for all other characteristics shown in Table 2.
Table 2.
Demographic characteristics of participants by recruitment cohort (n=960)
| Clinic N (%) |
Population N (%) |
P-valuea | |
|---|---|---|---|
| Age at index dateb | <0.001 | ||
| 18–<25 | 203 (46.1) | 288 (63.3) | |
| 25–<30 | 167 (38.0) | 144 (31.7) | |
| 30–35 | 70 (15.9) | 23 (5.1) | |
| Annual household income | <0.001 | ||
| Less than $50,000 | 112 (25.6) | 180 (37.3) | |
| $50,000–$99,999 | 247 (56.4) | 259 (53.7) | |
| Over $100,000 | 79 (18.0) | 43 (8.9) | |
| Education | 0.298 | ||
| Less than college graduate | 163 (35.7) | 195 (38.9) | |
| College graduate | 294 (64.3) | 306 (61.1) | |
| Race/ethnicity | 0.200 | ||
| White, non-Hispanic | 425 (92.6) | 474 (94.6) | |
| Hispanic or other non-white | 34 (7.4) | 27 (5.4) | |
| Religious Affiliation | 0.081 | ||
| Latter-day Saint | 338 (73.6) | 393 (78.4) | |
| Not Latter-day Saint | 121 (26.4) | 108 (21.6) | |
| Resident of Salt Lake Countyc | 0.004 | ||
| No | 271 (61.0) | 338 (70.1) | |
| Yes | 173 (39.0) | 144 (29.9) | |
| Body mass index | 0.276 | ||
| Underweight/Normal (≤25 kg/m2) | 220 (52.3) | 223 (48.6) | |
| Overweight/Obese (>25 kg/m2) | 201 (47.7) | 236 (51.4) | |
| Insurance coverage for fertility diagnosis or treatmentd | 0.784 | ||
| None or unsure | 32 (7.4) | 59 (13.6) | |
| At least some coverage | 100 (23.1) | 172 (39.6) | |
| Longest attempt to conceive | 0.068 | ||
| 12mo–<24months | 55 (12.0) | 84 (17.2) | |
| 24mo–<48months | 155 (33.8) | 154 (31.6) | |
| 48mo–<60months | 50 (10.9) | 65 (13.3) | |
| More than 60 months | 195 (42.5) | 185 (37.9) | |
| Time between start of first attempt to conceive and telephone interview |
|||
| <3 years | 26 (6.0) | 49 (11.3) | <0.001 |
| 3–<6 years | 119 (27.5) | 148 (34.2) | |
| 6–<10 years | 164 (37.9) | 206 (47.6) | |
| 10 years or more | 124 (28.6) | 30 (6.9) | |
| Most intensive treatmente | <0.001 | ||
| No medical treatment | 43 (9.4) | 207 (41.3) | |
| Ovulation drugs | 63 (13.7) | 128 (25.6) | |
| Intrauterine insemination | 142 (30.9) | 97 (19.4) | |
| In vitro fertilisation | 211 (46.0) | 69 (13.8) | |
| Subsequent pregnancyf | 0.334 | ||
| No | 172 (37.5) | 203 (40.5) | |
| Yes | 287 (62.5) | 298 (59.4) | |
| Subsequent live birthg | 0.390 | ||
| No | 209 (45.5) | 242 (48.3) | |
| Yes | 250 (54.5) | 259 (51.7) | |
| Total | 459 | 501 | |
P-value is for chi-square comparison of distribution between clinic and population cohorts.
Date of first clinic visit for clinic cohort; December 31, 2004 or December 31, 2008 for population cohort.
Salt Lake County is a predominantly urban county with the largest population of any county of the state. In 2008, 37% of the women ages 18–34 in Utah lived in Salt Lake County, https://ibis.health.utah.gov/query/result/pop/PopMain/Count.html, accessed December 28, 2015.
597 missing; this question was not asked for 504 women; for 93 women the question was asked but there was no response
as reported on screening questionnaire
as reported on the follow-up questionnaire
as identified in birth certificate linkage through December 31, 2010
Confirming eligibility criteria
After women were enrolled, we re-examined the eligibility criteria (determined previously on the screening questionnaire) based on women’s answers to the FEQ. As shown in Figure 1, among the clinical cohort, 94% of those eligible had all eligibility criteria that were determined on the screening questionnaire fully confirmed on the detailed FEQ. However, among the population group, only 61% fully confirmed. Among the 168 population-recruited women who did not have all eligibility criteria confirmed, the large majority (139 or 83%) had not been trying to conceive for a full 12 months at their index date (either December 31, 2004, or December 31, 2008), but of these, only 35 women had never tried to conceive for 12 months or more.
Assessment of Responders and Nonresponders
In Table 3, UCRM clinic-based women responding to the invitation to be screened are compared to those who did not respond to the invitation, based on the available data of age and subsequently linked live birth. (These data were not available for the RCC clinic-based sample.) Responders were slightly older, and were significantly more likely to have had a live birth, based on linkage to Utah birth certificates: 58% of responders versus 51% of nonresponders.
Table 3.
Characteristics of nonresponders and screened responders during clinic recruitmenta
| Nonresponders | Screened Responders | P-valueb | |
|---|---|---|---|
| Total | 1375 | 302 | |
| Age at recruitment | |||
| Continuous (mean yrs(SD)) | 37.2 (6.3) | 37.6 (5.7) | 0.344 |
| Age Category | N (%) | N (%) | |
| 18–24 | 29 (1.9) | 0(0) | 0.043 |
| 25–29 | 215 (13.8) | 30 (10.1) | |
| 30–34 | 478 (30.1) | 105 (35.4) | |
| 35–40 | 415 (26.7) | 80 (26.9) | |
| 40 or more | 416 (26.8) | 82 (27.6) | |
| Live birthc | |||
| No | 669 (48.7) | 126 (41.7) | 0.006 |
| Yes | 706 (51.3) | 176 (58.2) |
Includes only the Utah Center for Reproductive Medicine. Does not include the Reproductive Care Center.
P-value is for t-test for continuous age and chi-square comparison of distribution between nonresponders and responders.
Based on linkage to Utah birth certificates through December 31, 2010.
In Table 4, population-based women responding to the invitation to be screened are compared to those who did not respond to the invitation, based on the available data of age, race/ethnicity, education, and subsequently linked live birth or adoption. In both 2004-indexed and 2008-indexed cohorts, responders exhibit a higher level of education, and are more likely to be linked to a subsequent birth.
Table 4.
Characteristics of nonresponders and responders during population recruitment
| Index date 2004b | Index date 2008c | |||||
|---|---|---|---|---|---|---|
| Non- responders n (%) |
Responders n(%) |
P-value | Non- responders n(%) |
Responders n(%) |
P-value | |
| Total | 4332 | 1491 | 1737 | 1377 | ||
| Age at recruitment | ||||||
| 20–24 | 0 (0) | 0 (0) | 0.02 | 52 (3.0) | 55 (4.0) | 0.214 |
| 25–29 | 1372 (31.7) | 413 (27.7) | 1112(64.1) | 881 (63.9) | ||
| 30–39 | 2937 (67.8) | 1069 (71.7) | 573 (32.9) | 441 (33.0) | ||
| 40–44 | 23 (0.1) | 9 (0.6) | 0 (0) | 0 (0) | ||
| Race | ||||||
| White | 4118 (95.1) | 1441 (96.6) | 0.01 | 1650 (94.9) | 1308 (95.0) | 0.947 |
| Non-white | 211 (4.9) | 50 (3.4) | 87 (5.1) | 69 (5.1) | ||
| Unknown | 3 (0.0) | 0 | 0 (0.0) | 0 (0.0) | ||
| Ethnicity | ||||||
| Hispanic | 149 (3.4) | 26 (1.7) | <0.001 | 52 (3.0) | 41 (2.9) | 0.108 |
| Non-Hispanic | 2582 (59.6) | 979 (65.7) | 556 (32.1) | 551(40.1) | ||
| Unknown | 1601 (37.0) | 486 (32.6) | 1129 (65.0) | 785 (57.0) | ||
| Educationd | ||||||
| Less than high school | 34 (0.8) | 7 (0.5) | <0.001 | 17 (1.0) | 0 (0.0) | <0.05 |
| High school graduate | 600 (13.9) | 137 (9.2) | 69 (3.9) | 41 (3.0) | ||
| Some college | 979 (22.6) | 339 (22.7) | 278 (16.0) | 207 (15.0) | ||
| College graduate | 1364 (31.5) | 646 (43.3) | 226 (13.0) | 317 (23.0) | ||
| Unknown | 1355 (31.3) | 362 (24.3) | 1164 (67.1) | 799 (58.0) | ||
| Childbearing in Utahe | ||||||
| No births | 1335 (30.8) | 362 (24.3) | <0.001 | 1164 (67.0) | 785 (57.0) | <0.01 |
| At least one birth | 2997 (69.2) | 1129 (75.7) | 573 (33.1) | 592 (42.9) | ||
| At least one adoptionf | 93 (2.1) | 43 (2.9) | 17 (0.9) | 0 (0.0) | ||
P-value is for chi-square comparison of distribution between nonresponders and responders.
For the sample with the December 31, 2004 index date, there were 7400 letters originally sent, among which were 1577 incorrect postal addresses, not included in the table.
For the sample with the December 31, 2008 index date, there were 7857 letters originally sent, among which were 4743 incorrect postal addresses, not included in the table.
At time of first birth, where applicable.
Based on linkage to Utah birth certificates and adoption certificates through December 31, 2010.
May have also had a birth.
Comment
We successfully recruited parallel retrospective cohorts of women for this study: a clinic-based cohort from the two major specialty fertility clinics in Utah during the study time period (n=459), and a cohort from population-based sampling (n=501). The two cohorts were similar with respect to race/ethnicity, education, and religious affiliation. Women recruited through the clinic were older and had higher annual household income (Table 2). After the index date, women in the clinic-based cohort were more likely to have received intrauterine insemination or in vitro fertilization. Population-cohort women were more likely to have received no medical treatment, or ovulation drugs, which may often be prescribed by obstetrician-gynecologists or primary care clinicians.17
The recruitment letter for the clinics referred to “treatment choices, outcomes, and costs,” while the letter for the population cohort referred to “times in your life you may have tried to get pregnant, and care you have received, if any.” This language was designed for flexibility with regard to the population cohort, who may or may not have considered or obtained any treatment. However, it is possible that the difference in wording may have had some differential influence on women’s responses.
While the final eligibility criteria were the same for both cohorts, the preliminary information available for initial sampling differed between cohorts (Table 1). For preliminary sampling of the population cohort, we picked a marriage date that was 2–5 years prior to the index date, to capture a broad spectrum of prior time trying to conceive. This may have excluded some couples with longer times attempting to conceive (correlated with earlier marriage dates), as well as longer times between onset of time trying to conceive and the study interview, relative to the clinic cohort. In addition, the population cohort was restricted by sampling to married women. Although we do not have marriage information from the clinics, anecdotally the very large majority of the heterosexual couples seen at the clinics during this time frame were known to be married.
After enrollment, we used the FEQ to verify the eligibility criteria assessed on the screening questionnaire and found a high level of confirmation in the clinic cohort (94%), but a lower level in the population cohort (61%), due mostly to discrepancies in when the women had first reached 12 months of trying to conceive. Women recruited from the clinics had a more memorable index date based on their first clinic visit; this likely facilitated a stronger recall of this criterion. If treatment becomes more (or less) available over time, this could influence treatment choices for some of the population-based women who first reached 12 months of trying to conceive somewhat later than the clinic women.
We cannot rule out possible selection bias through nonresponse, because we don’t know what proportion of nonresponders would have been eligible for the study. However, there was evidence suggesting selection bias at the level of initial recruiting: women in both clinic and population cohorts who had a live birth were more likely to respond. This suggests that studies that follow-up clinic patients after failed fertility treatment based solely on re-contacting them will have a higher proportion women who conceived among their responders.18 In our study, responders were also slightly older in the clinic and in one of the population subgroups, and more educated (assessed in the population cohort only). Buck Louis and colleagues recently reported a new population-based cohort of births, about one fourth of whom received fertility treatment, based on sampling through birth certificates in New York State, and recruiting through mail and telephone.9 They reported that women who participated had a higher educational attainment than those who did not participate, similar to our study. Our study differs in selecting only women with subfertility, and including women who did not have a live birth.
Declerq and colleagues have linked information from Massachusetts birth certificates to information from databases on statewide hospitalization and assisted reproductive technology treatment to identify births resulting from IVF, subfertility without IVF treatment related to the specific birth based on prior IVF or hospital admissions, and births with no record of any kind of fertility treatment.19, 20 This approach can efficiently identify large numbers of births and does not suffer from the same kind of respondent bias, but it still does not identify subfertile women who conceive without any medical treatment, nor does it identify subfertile women who do not have a live birth. If birth certificates were to include a universal simple retrospective question about time to pregnancy, it would be possible to identify the substantial proportion of subfertile women who conceive without medical treatment.21
In research on human infertility, many studies have focused on ART.22–24 Others have looked at a spectrum of treatments, but they overlook many women or couples who are subfertile but who do not seek specialty medical treatment.25–29 Our results confirm that couples seeking care at specialty clinics are not representative of all couples experiencing subfertility, in terms of demographics, reproductive and treatment history, or outcomes.
Because our study aimed to compare cumulative pregnancy rates over time with women who were eligible for different treatments or received no medical treatment, we enrolled women age 35 or younger at the index date. Thus, our sample does not reflect the full age spectrum of those seeking specialty medical care for fertility services.
Insurance coverage for fertility services differs between states and nations and has an impact on seeking services.30 Caution is needed in extrapolating from our specific results to other states or countries.
In the absence of randomized trials, parallel data from comparable clinic-based and population-based cohorts can identify subfertile couples who have received a spectrum of treatments or no treatment, which will contribute to understanding the relative contribution of underlying conditions and treatments to pregnancy rates and the health of the newborns.31,32 Such information is critical for couples facing decisions about fertility treatment, and for providers seeking to care for these couples and their children.
Conclusions
Women recruited through clinic-based sampling for retrospective assessment of subfertility, treatments, and outcomes are older, have higher income, and receive more invasive treatment than women recruited through population-based sampling, who are more likely to have received less invasive treatment, or no medical treatment. However, in both clinic-based and population-based sampling, women who have had a live birth are more likely to respond to retrospective recruitment.
Supplementary Material
Acknowledgments
This study was funded by the Eunice Kennedy Shriver National Institute of Child Health and Human Development, R21 HD060213-01A1. We thank Scarlett Reeves, Mandy Ward, Stephanie Croasdell, Leslie McNaughtan, Alicia Fadrowsky, Melissa Sperry, Alessandra Zimmerman, Varsha Iyer, Erik Linn, Melody Anderson, Namealoha Sells, Rita Sharshiner, Mohammed Al-Temimi, Jahn Barlow (RGE), Alison Fraser (UPDB), Marcia Feldkamp (Utah Birth Defects Network), who all made valuable contributions to study design or execution. We thank physicians at the two clinics, particularly C. Matthew Peterson, Kirtly P. Jones, Shawn Gurtcheff, James Heiner, and Keith Blauer, as well as Carra Christopher at UCRM and administrative staff at RCC. We thank Mary Croughan, who served as a consultant for this study. We acknowledge the support of the Huntsman Cancer Institute for datasets within the Utah Population Database.
Funding: This work was supported by the Eunice Kennedy Shriver National Institute of Child Health and Human Development, R21 HD060213-01A1, and a grant from the Primary Children’s Medical Center Foundation, Salt Lake City, Utah.
Footnotes
Conflict of Interest: The authors declare no conflict of interest.
IRB: University of Utah IRB #27783 and Utah Department of Health IRB #199
References
- 1.American Society for Reproductive Medicine. Definitions of infertility and recurrent pregnancy loss. Fertility and Sterility. 2008;89:1603. doi: 10.1016/j.fertnstert.2008.03.002. [DOI] [PubMed] [Google Scholar]
- 2.Evers JL. Female subfertility. Lancet. 2002;360:151–159. doi: 10.1016/S0140-6736(02)09417-5. [DOI] [PubMed] [Google Scholar]
- 3.Oakley L, Doyle P, Maconochie N. Lifetime prevalence of infertility and infertility treatment in the UK: results from a population-based survey of reproduction. Human Reproduction. 2008;23:447–450. doi: 10.1093/humrep/dem369. [DOI] [PubMed] [Google Scholar]
- 4.Thoma ME, McLain AC, Louis JF, King RB, Trumble AC, Sundaram R, et al. Prevalence of infertility in the United States as estimated by the current duration approach and a traditional constructed approach. Fertility and Sterility. 2013;99:1324–1331. e1321. doi: 10.1016/j.fertnstert.2012.11.037. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Kessler LM, Craig BM, Plosker SM, Reed DR, Quinn GP. Infertility evaluation and treatment among women in the United States. Fertility and Sterility. 2013;100:1025–1032. e1022. doi: 10.1016/j.fertnstert.2013.05.040. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Smith JF, Eisenberg ML, Glidden D, Millstein SG, Cedars M, Walsh TJ, et al. Socioeconomic disparities in the use and success of fertility treatments: analysis of data from a prospective cohort in the United States. Fertility and Sterility. 2011;96:95–101. doi: 10.1016/j.fertnstert.2011.04.054. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Simonsen SE, Baksh L, Stanford JB. Infertility treatment in a population-based sample: 2004–2005. Maternal Child Health Journal. 2012;16:877–886. doi: 10.1007/s10995-011-0809-6. [DOI] [PubMed] [Google Scholar]
- 8.Johnson NP, Proctor M, Farquhar CM. Gaps in the evidence for fertility treatment-an analysis of the Cochrane Menstrual Disorders and Subfertility Group database. Human Reproduction. 2003;18:947–954. [PubMed] [Google Scholar]
- 9.Buck Louis GM, Hediger ML, Bell EM, Kus CA, Sundaram R, McLain AC, et al. Methodology for establishing a population-based birth cohort focusing on couple fertility and children's development, the Upstate KIDS Study. Paediatric and Perinatal Epidemiology. 2014;28:191–202. doi: 10.1111/ppe.12121. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Romundstad LB, Romundstad PR, Sunde A, Düring Vv, Skjaerven R, Gunnell D, et al. Effects of technology or maternal factors on perinatal outcome after assisted fertilisation: a population-based cohort study. Lancet. 2008;372:737–743. doi: 10.1016/S0140-6736(08)61041-7. [DOI] [PubMed] [Google Scholar]
- 11.Stanford J, Simonsen S, Baksh L. Fertility treatments and adverse perinatal outcomes in a population-based sampling of births in Florida, Maryland, and Utah: a cross-sectional study. British Journal of Obstetrics and Gynecology. 2015 doi: 10.1111/1471-0528.13510. [DOI] [PubMed] [Google Scholar]
- 12.Lynch CD, Louis GM, Lahti MC, Pekow PS, Nasca PC, Cohen B. The birth certificate as an efficient means of identifying children conceived with the help of infertility treatment. American Journal of Epidemiology. 2011;174:211–218. doi: 10.1093/aje/kwr036. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Buck Louis GM, Schisterman EF, Dukic VM, Schieve LA. Research hurdles complicating the analysis of infertility treatment and child health. Human Reproduction. 2005;20:12–18. doi: 10.1093/humrep/deh542. [DOI] [PubMed] [Google Scholar]
- 14.Stanford JB, Mikolajczyk RT, Lynch CD, Simonsen SE. Cumulative pregnancy probabilities among couples with subfertility: effects of varying treatments. Fertility and Sterility. 2010;93:2175–2181. doi: 10.1016/j.fertnstert.2009.01.080. [DOI] [PubMed] [Google Scholar]
- 15.Resource for genetic and epidemiologic research. University of Utah; [accessed January 14, 2015]. http://rge.utah.edu/ [Google Scholar]
- 16.Thomas FS, Stanford JB, Sanders JN, Gurtcheff SE, Gibson M, Porucznik CA, et al. Development and initial validation of a fertility experiences questionnaire. Reproductive Health. 2015;12:62. doi: 10.1186/s12978-015-0054-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Stanford JB, Martin JC, Gibson M, Birdsall E, Brixner DI. Use of clomiphene citrate in the University of Utah Community Clinics. Journal of Reproductive Medicine. 2013;58:229–233. [PubMed] [Google Scholar]
- 18.Troude P, Bailly E, Guibert J, Bouyer J, de la Rochebrochard E, Group D. Spontaneous pregnancies among couples previously treated by in vitro fertilization. Fertility and Sterility l. 2012;98:63–68. doi: 10.1016/j.fertnstert.2012.03.058. [DOI] [PubMed] [Google Scholar]
- 19.Declercq ER, Belanoff C, Diop H, Gopal D, Hornstein MD, Kotelchuck M, et al. Identifying women with indicators of subfertility in a statewide population database: operationalizing the missing link in assisted reproductive technology research. Fertility and Sterility. 2014;101:463–471. doi: 10.1016/j.fertnstert.2013.10.028. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Declercq E, Luke B, Belanoff C, Cabral H, Diop H, Gopal D, et al. Perinatal outcomes associated with assisted reproductive technology: the Massachusetts Outcomes Study of Assisted Reproductive Technologies (MOSART) Fertility and Sterility. 2015 doi: 10.1016/j.fertnstert.2014.12.119. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Joffe M. Validity of self-reported time to pregnancy. Epidemiology. 2010;21:160–161. doi: 10.1097/EDE.0b013e3181c1ec69. author reply 161. [DOI] [PubMed] [Google Scholar]
- 22.Sunderam S, Kissin DM, Crawford SB, Folger SG, Jamieson DJ, Warner L, et al. Assisted Reproductive Technology Surveillance - United States, 2012. Morbidity Mortalility Weekly Review Surveillance Summary. 2015;64:1–29. [PubMed] [Google Scholar]
- 23.Henningsen AA, Gissler M, Skjaerven R, Bergh C, Tiitinen A, Romundstad LB, et al. Trends in perinatal health after assisted reproduction: a Nordic study from the CoNARTaS group. Human Reproduction. 2015;30:710–716. doi: 10.1093/humrep/deu345. [DOI] [PubMed] [Google Scholar]
- 24.Xiong X, Dickey RP, Pridjian G, Buekens P. Maternal age and preterm births in singleton and twin pregnancies conceived by in vitro fertilisation in the United States. Paediatric and Perinatal Epidemiology. 2015;29:22–30. doi: 10.1111/ppe.12166. [DOI] [PubMed] [Google Scholar]
- 25.Hull MG, Glazener CM, Kelly NJ, Conway DI, Foster PA, Hinton RA, et al. Population study of causes, treatment, and outcome of infertility. British Medical Journal (Clinical Research Edition) 1985;291:1693–1697. doi: 10.1136/bmj.291.6510.1693. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Haadsma ML, Groen H, Fidler V, Bukman A, Roeloffzen EM, Groenewoud ER, et al. The predictive value of ovarian reserve tests for spontaneous pregnancy in subfertile ovulatory women. Human Reproduction. 2008;23:1800–1807. doi: 10.1093/humrep/den234. [DOI] [PubMed] [Google Scholar]
- 27.Snick HK, Snick TS, Evers JL, Collins JA. The spontaneous pregnancy prognosis in untreated subfertile couples: the Walcheren primary care study. Human Reproduction. 1997;12:1582–1588. doi: 10.1093/humrep/12.7.1582. [DOI] [PubMed] [Google Scholar]
- 28.Smith JF, Eisenberg ML, Millstein SG, Nachtigall RD, Sadetsky N, Cedars MI, et al. Fertility treatments and outcomes among couples seeking fertility care: data from a prospective fertility cohort in the United States. Fertility and Sterility. 2011;95:79–84. doi: 10.1016/j.fertnstert.2010.06.043. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Messerlian C, Platt RW, Tan SL, Gagnon R, Basso O. Low-technology assisted reproduction and the risk of preterm birth in a hospital-based cohort. Fertility and Sterility. 2015;103:81–88. e82. doi: 10.1016/j.fertnstert.2014.10.006. [DOI] [PubMed] [Google Scholar]
- 30.Hammoud AO, Gibson M, Stanford J, White G, Carrell DT, Peterson M. In vitro fertilization availability and utilization in the United States: a study of demographic, social, and economic factors. Fertility and Sterility. 2009;91:1630–1635. doi: 10.1016/j.fertnstert.2007.10.038. [DOI] [PubMed] [Google Scholar]
- 31.Jackson RA, Gibson KA, Wu YW, Croughan MS. Perinatal outcomes in singletons following in vitro fertilization: a meta-analysis. Obstetrics and Gynecology. 2004;103:551–563. doi: 10.1097/01.AOG.0000114989.84822.51. [DOI] [PubMed] [Google Scholar]
- 32.Pinborg A, Wennerholm UB, Romundstad LB, Loft A, Aittomaki K, Söderström-Anttila V, et al. Why do singletons conceived after assisted reproduction technology have adverse perinatal outcome? Systematic review and meta-analysis. Human Reproduction Update. 2013;19:87–104. doi: 10.1093/humupd/dms044. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.